Hexapawn: A Game You Play to Lose

New York: IBM, 1969. Vintage instructions detailing the rules and uses of Hexapawn as they apply to machine learning and programming. Scarce, with only one copy in OCLC.

Hexapawn was invented by mathematician Martin Gardner, and the rules were first published in in 1962. The games was adopted by IBM as a method of teaching computers to play by a process of eliminating moves that lead to defeat through multiple play throughs. An early example of the heuristic approach to artificial intelligence and machine learning, similar to that used to train machines to play more complex games like chess.

13.5 x 14.5 inches on card stock. Two leaves, with printing on rectos and versos. Very Good plus, with light soil to the first and last pages. In the original IBM envelope, Very Good with some splitting.

[Book #149002]

Price: $950.00